主要研究降落图像的天体表面障碍物的识别和着陆区域的选取技术.首先应用数学形态学中的预、底帽变换以及腐蚀、膨胀等算法对探测目标表面的特征信息进行识别,通过图像滤噪、特征提取,得出障碍信息;然后对提取出的障碍信息进行sobel边缘检测;最后提出一种改进的霍夫变换圆形检测算法求出最大的安全着陆区域.对实拍序列降落图像进行系统仿真以及对比实验.结果表明,此种方法可以准确、快速地为探测器提供安全着陆信息,并能够有效地为探测器提供出备选的安全着陆区域以及最佳着陆点.
This paper mainly focuses on the investigation about the recognition of the aster' s surface obstructions and the selection of landing sites with the help from landing imaging. Firstly, by using mathematical morphology of top and bottom hat transformations, cor- rosion and expansion algorithm, the denoising feature information of the target surface can be obtained; and then the edges on the ex- tracted feature information can be detected using Sobel algorithm. Finally, a circular detection algorithm based on an improved Hough transform is introduced to obtain the largest safe landing site. Tile simulation of the actual shooting sequence landing images demon- strates that these methods can provide the safety information on the landing sites exactly and fleetly.